The authors argue that generative AI introduces a new class of alignment risks because interaction itself becomes a mechanism of influence. Humans adapt their behavior in response to AI outputs, ...
AI can act as a second set of eyes for content, helping evaluate intent signals, compare top results, and refocus ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback